I. Introduction
Computer vision has become an interdisciplinary subject integrating digital image processing, machine learning, computer science and many other categories. Its application in sports is becoming more and more extensive and in-depth. Martial arts require high quality for athletes' postures and actions. In some vigorous exercise scenes, the duration of actions is short, it is very difficult for human eyes to evaluate the action, and it is difficult to avoid the subjectivity and errors. If a camera can be used to collect body motion, the joint point position, angle and other important parameters of human action can be obtained through the processing of video frames, which aim to analyze and judge the motion and give more scientific guidance for users. Thus, computer vision brings great benefits to martial arts which include more scientific training, more effective technical and tactical analysis and so on. Although the application of computer vision in martial arts started early, it has not been well summarized. Obviously, a systematic review of these applications is necessary.